Capital Bancorp Inc (CBNK) stock intrinsic value fair value estimate & owner earnings analysis

Capital Bancorp is the Rockville, Maryland holding company for Capital Bank, N.A., about $3.9B in assets and $422M of equity at June 30, 2026. It has four businesses. The Commercial Bank lends to businesses, real-estate investors and not-for-profits, mainly around Washington, D.C. and Baltimore, and gathers deposits nationally from groups such as homeowners' associations and mortgage companies. OpenSky is a nationwide digital credit-card business offering secured and, to existing secured-card customers, unsecured cards. Windsor Advantage runs an SBA 7(a) and USDA lending platform for other community banks and credit unions. Capital Bank Home Loans makes residential mortgages for sale. The late-2024 purchase of Integrated Financial Holdings brought in Windsor and a book of commercial loans, so we measure the business from 2025, its first full year in its current shape. What changed since our last look. We previously read the first half of 2026 as confirming the 2025 earnings level. That was wrong: the one-time $3.5M after-tax gain from calling acquired brokered deposits came in the third quarter of 2025, not the first half, so the like-for-like comparison is weaker than we thought. The quarterly segment detail also shows OpenSky's pre-tax profit falling to $0.9M in the first quarter of 2026 and $1.0M in the second, against $3.1M and $2.8M a year earlier and $12.6M for 2025 as a whole (excluding a $2.0M one-time credit from selling charged-off card debt in the fourth quarter). The company attributes this to higher professional fees and headcount for the unsecured and legacy card products, faster depreciation of OpenSky technology, lower unsecured card fees and reserves built for unsecured balance growth. Two consecutive quarters is enough for us to treat this as the current level rather than a blip, so our base case starts lower and our value falls from $82 to $78. The margin-of-safety price is also lower, from $45 to $28, for the reasons set out at the end. How we measure earnings. For a bank, operating cash flow is distorted by mortgages that are made and sold within weeks, so we treat those flows as timing and keep the profit on each sale in earnings. We count the provision for loan losses as a real, recurring cost, and deduct stock-based pay and rent. On that basis earnings per share were $0.60 in 2017, $1.01 (2018), $1.23 (2019), $1.72 (2020), $2.75 (2021, a mortgage-refinance boom), $2.89 (2022), $2.37 (2023), $1.58 (2024, held down by acquisition costs, a day-one reserve on acquired loans and an investment write-down) and $3.50 (2025). Removing the $3.5M deposit-call gain and adding back $2.6M of after-tax merger costs, 2025 earnings were about $57.8M, or $3.53 per share on the 16.4M diluted shares now outstanding. Purchase-accounting income on the acquired loans added only 0.04 points to loan yields, so it does not flatter this figure. The starting point. Our year-one estimate is $3.26 per share, about 7% below 2025. The Commercial Bank's pre-tax profit rose 7% in the first half of 2026 ($30.9M against $28.8M) and Windsor's rose to $3.7M from $2.1M, but OpenSky fell to $1.9M from $5.9M and the mortgage unit lost $2.2M against $1.4M. First-half earnings per share were flat at $1.60, and we do not assume a faster pace in year one than the business is actually showing. Growth and the capital it needs. A bank can only grow loans if it keeps capital behind them, so wherever we credit growth we hold back the retained earnings it needs. The growth we credit is the company's own record: portfolio loans grew about 14% a year from 2018 to 2023, funded by retained earnings and deposits through the 2020 pandemic and the 2022-23 rate shock without selling new shares, and grew another 12-13% in 2025 and 12.6% in the year to June 2026. We credit well under half that pace and taper it to zero over years 11-20, then hold earnings flat with no further growth, relying on a deposit and commercial-lending franchise the bank has run since 1999. All values use a 6% discount rate. Because the company's borrowings are part of its raw material as a lender, we value earnings after interest and do not deduct deposits or its small remaining $2.1M of other borrowings. Base case, $78 per share. Earnings start at $3.26 and grow 6% a year through year 10, earning about 13% on the capital retained to fund that growth, so a little under half of each year's earnings is kept in the bank. OpenSky stays at its current reduced profit of roughly $4M pre-tax a year; we do not credit a recovery that has not happened. We deduct about $0.15 per share of extra loan losses in year two from the rise in problem loans described below. With no growth at all, the same starting earnings would be worth about $54. Bear case, $46 per share. Four named risks play out. First, OpenSky, whose profit has already fallen by roughly 70%, goes to break-even. Second, a credit downturn: nonperforming assets rose to $60.8M (1.56% of assets, 1.85% of loans) from $36.1M a year earlier, including a single $9.7M legacy relationship moved to nonaccrual in the second quarter, and substandard loans stand at $70.6M. We assume net charge-offs rise from about 0.45% of loans to 1.2% for two years, costing about $1.08 per share each year, and that a similar downturn recurs a decade later. Third, margins tighten: the Commercial Bank's loan yield fell 0.11 points in 2025 on lower rates, and in the 2023 rate shock our measure of earnings per share fell 18%. We take 10% off earnings for three years and leave a permanent 5% reduction after that. Fourth, no growth is credited. The bank stays solidly profitable at about $2.90 per share in normal years, and its $422M of equity and $54M loan-loss reserve comfortably absorb these losses. The bear case is a lower-earning bank, not an impaired one. It is higher than our previous $43 because OpenSky's decline is now already in the starting level rather than a future loss, and because uninsured deposits fell to about a third of deposits from 41% at year-end, so we no longer charge a permanent funding-cost penalty for them. Bull case, $106 per share. The bank's own levers keep working. Earnings start at $3.30, reflecting the stronger second half the bank has shown seasonally, and grow 9% a year through year 10, still below the loan growth it has actually delivered, earning 14% on retained capital. OpenSky's profit recovers to about $10M pre-tax as unsecured balances season past the reserves built up front when they are made, still below the $12.6M to $17.3M it earned in 2024 and 2025. Windsor keeps its current pace. We do not credit a return of mortgage-boom profits or any widening of margins. What could make this wrong. The bank is funded roughly 9-to-1 by deposits, so credit losses beyond the bear case cut straight into equity, and problem loans are rising. OpenSky's heavy spending may not produce a return, and the mortgage unit has lost money for three years. Brokered deposits are about 12% of the total. Segment figures for 2026 come from the company's quarterly earnings releases; we could not review the full second-quarter report text. The margin-of-safety price is $28, about 39% below the bear case and about 64% below the base case. We set the discount this wide for three reasons. First, the bear case is itself a stressed outcome we could be wrong about, and a price at the bear case would leave nothing for error in our credit or OpenSky assumptions. Second, a lender funded roughly 9-to-1 by deposits can lose a large share of its equity to a small misjudgement in credit. Third, the estimate should not depend on the market paying an unusual price later. From $28, a value of about 13 times earnings three years out (about $48), plus dividends, implies roughly 20% a year. The previous margin-of-safety price of $45 sat almost exactly at the bear case, and for it to imply that return the shares would have needed to reach about 21 times earnings within three years, which small banks rarely command.